ABSTRACT While considerable evidence demonstrates that CD8+ TILs reactive to mutated or non-mutated tumor antigens play an important role in anti-tumor immunity, expanding evidence indicates that bona fide tumor-reactive T cells comprise only a minor fraction of all TILs in many human tumors, indicating that most CD8+ TILs have undefined specificity. Based on this long-standing question, we examined the hypothesis that a substantial fraction of TILs may represent CD8+ memory-phenotype T cells (CD8-MP cells), a unique population of cells of unknown antigen specificity that comprise 5-10% of CD8+ T cells in unprimed mice, exhibit common hallmarks of prior antigen experience, and have the capacity to rapidly expand and produce IFN-? during an immune response. In preliminary work, we found that CD8-MP cells make substantial contributions to the immune infiltrate of oncogene-driven prostate tumors, and express high densities of the PD-1 inhibitory receptor. Given these unique insights, we embarked on parallel studies aimed at further elucidating the fundamental biology of CD8-MP cells, using a clonal approach to define the developmental trajectories of these cells. Our new data reveal that the differentiation of many CD8-MP clones is triggered by recognition of self-ligands in the thymus via a reproducible, orchestrated process, challenging current thought suggesting that CD8-MP cells differentiate in the periphery in response to homeostatic signals. In Aim 1, we will define the function of CD8-MP cells in anti-tumor immunity, testing the hypothesis that self-ligand recognition drives the early entry of CD8-MP cells into developing tumors, and that CD8-MP cells enhance anti-tumor immunity by catalyzing broader immune cell infiltration. In addition, we will identify novel markers that can be used to identify intratumoral CD8-MP cells, thereby enabling the broader study of CD8-MP cells in murine cancer models and human cancer patients. In Aim 2, we will elucidate the molecular and cellular mechanisms that direct CD8-MP differentiation, testing the hypothesis that CD8-MP differentiation is a two-step process triggered by TCR-dependent recognition of self-ligands presented by classical dendritic cells in the thymus. We will also utilize a unique T cell antigen discovery assay to identify natural self-peptides recognized by CD8-MP cells, thereby opening new areas of inquiry that were previously inaccessible. Ultimately, defining the function of intratumoral CD8-MP cells and the blueprints of CD8-MP differentiation in mice is expected to open new avenues for the study and manipulation of these cells in humans.